fs/btrfs/backref.c

   1 /*
   2  * Copyright (C) 2011 STRATO.  All rights reserved.
   3  *
   4  * This program is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU General Public
   6  * License v2 as published by the Free Software Foundation.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  11  * General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public
  14  * License along with this program; if not, write to the
  15  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  16  * Boston, MA 021110-1307, USA.
  17  */
  18
  19 #include <linux/vmalloc.h>
  20 #include "ctree.h"
  21 #include "disk-io.h"
  22 #include "backref.h"
  23 #include "ulist.h"
  24 #include "transaction.h"
  25 #include "delayed-ref.h"
  26 #include "locking.h"
  27
  28 /* Just an arbitrary number so we can be sure this happened */
  29 #define BACKREF_FOUND_SHARED 6
  30
  31 struct extent_inode_elem {
  32         u64 inum;
  33         u64 offset;
  34         struct extent_inode_elem *next;
  35 };
  36
  37 static int check_extent_in_eb(struct btrfs_key *key, struct extent_buffer *eb,
  38                                 struct btrfs_file_extent_item *fi,
  39                                 u64 extent_item_pos,
  40                                 struct extent_inode_elem **eie)
  41 {
  42         u64 offset = 0;
  43         struct extent_inode_elem *e;
  44
  45         if (!btrfs_file_extent_compression(eb, fi) &&
  46             !btrfs_file_extent_encryption(eb, fi) &&
  47             !btrfs_file_extent_other_encoding(eb, fi)) {
  48                 u64 data_offset;
  49                 u64 data_len;
  50
  51                 data_offset = btrfs_file_extent_offset(eb, fi);
  52                 data_len = btrfs_file_extent_num_bytes(eb, fi);
  53
  54                 if (extent_item_pos < data_offset ||
  55                     extent_item_pos >= data_offset + data_len)
  56                         return 1;
  57                 offset = extent_item_pos - data_offset;
  58         }
  59
  60         e = kmalloc(sizeof(*e), GFP_NOFS);
  61         if (!e)
  62                 return -ENOMEM;
  63
  64         e->next = *eie;
  65         e->inum = key->objectid;
  66         e->offset = key->offset + offset;
  67         *eie = e;
  68
  69         return 0;
  70 }
  71
  72 static void free_inode_elem_list(struct extent_inode_elem *eie)
  73 {
  74         struct extent_inode_elem *eie_next;
  75
  76         for (; eie; eie = eie_next) {
  77                 eie_next = eie->next;
  78                 kfree(eie);
  79         }
  80 }
  81
  82 static int find_extent_in_eb(struct extent_buffer *eb, u64 wanted_disk_byte,
  83                                 u64 extent_item_pos,
  84                                 struct extent_inode_elem **eie)
  85 {
  86         u64 disk_byte;
  87         struct btrfs_key key;
  88         struct btrfs_file_extent_item *fi;
  89         int slot;
  90         int nritems;
  91         int extent_type;
  92         int ret;
  93
  94         /*
  95          * from the shared data ref, we only have the leaf but we need
  96          * the key. thus, we must look into all items and see that we
  97          * find one (some) with a reference to our extent item.
  98          */
  99         nritems = btrfs_header_nritems(eb);
 100         for (slot = 0; slot < nritems; ++slot) {
 101                 btrfs_item_key_to_cpu(eb, &key, slot);
 102                 if (key.type != BTRFS_EXTENT_DATA_KEY)
 103                         continue;
 104                 fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
 105                 extent_type = btrfs_file_extent_type(eb, fi);
 106                 if (extent_type == BTRFS_FILE_EXTENT_INLINE)
 107                         continue;
 108                 /* don't skip BTRFS_FILE_EXTENT_PREALLOC, we can handle that */
 109                 disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
 110                 if (disk_byte != wanted_disk_byte)
 111                         continue;
 112
 113                 ret = check_extent_in_eb(&key, eb, fi, extent_item_pos, eie);
 114                 if (ret < 0)
 115                         return ret;
 116         }
 117
 118         return 0;
 119 }
 120
 121 /*
 122  * this structure records all encountered refs on the way up to the root
 123  */
 124 struct __prelim_ref {
 125         struct list_head list;
 126         u64 root_id;
 127         struct btrfs_key key_for_search;
 128         int level;
 129         int count;
 130         struct extent_inode_elem *inode_list;
 131         u64 parent;
 132         u64 wanted_disk_byte;
 133 };
 134
 135 static struct kmem_cache *btrfs_prelim_ref_cache;
 136
 137 int __init btrfs_prelim_ref_init(void)
 138 {
 139         btrfs_prelim_ref_cache = kmem_cache_create("btrfs_prelim_ref",
 140                                         sizeof(struct __prelim_ref),
 141                                         0,
 142                                         SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
 143                                         NULL);
 144         if (!btrfs_prelim_ref_cache)
 145                 return -ENOMEM;
 146         return 0;
 147 }
 148
 149 void btrfs_prelim_ref_exit(void)
 150 {
 151         if (btrfs_prelim_ref_cache)
 152                 kmem_cache_destroy(btrfs_prelim_ref_cache);
 153 }
 154
 155 /*
 156  * the rules for all callers of this function are:
 157  * - obtaining the parent is the goal
 158  * - if you add a key, you must know that it is a correct key
 159  * - if you cannot add the parent or a correct key, then we will look into the
 160  *   block later to set a correct key
 161  *
 162  * delayed refs
 163  * ============
 164  *        backref type | shared | indirect | shared | indirect
 165  * information         |   tree |     tree |   data |     data
 166  * --------------------+--------+----------+--------+----------
 167  *      parent logical |    y   |     -    |    -   |     -
 168  *      key to resolve |    -   |     y    |    y   |     y
 169  *  tree block logical |    -   |     -    |    -   |     -
 170  *  root for resolving |    y   |     y    |    y   |     y
 171  *
 172  * - column 1:       we've the parent -> done
 173  * - column 2, 3, 4: we use the key to find the parent
 174  *
 175  * on disk refs (inline or keyed)
 176  * ==============================
 177  *        backref type | shared | indirect | shared | indirect
 178  * information         |   tree |     tree |   data |     data
 179  * --------------------+--------+----------+--------+----------
 180  *      parent logical |    y   |     -    |    y   |     -
 181  *      key to resolve |    -   |     -    |    -   |     y
 182  *  tree block logical |    y   |     y    |    y   |     y
 183  *  root for resolving |    -   |     y    |    y   |     y
 184  *
 185  * - column 1, 3: we've the parent -> done
 186  * - column 2:    we take the first key from the block to find the parent
 187  *                (see __add_missing_keys)
 188  * - column 4:    we use the key to find the parent
 189  *
 190  * additional information that's available but not required to find the parent
 191  * block might help in merging entries to gain some speed.
 192  */
 193
 194 static int __add_prelim_ref(struct list_head *head, u64 root_id,
 195                             struct btrfs_key *key, int level,
 196                             u64 parent, u64 wanted_disk_byte, int count,
 197                             gfp_t gfp_mask)
 198 {
 199         struct __prelim_ref *ref;
 200
 201         if (root_id == BTRFS_DATA_RELOC_TREE_OBJECTID)
 202                 return 0;
 203
 204         ref = kmem_cache_alloc(btrfs_prelim_ref_cache, gfp_mask);
 205         if (!ref)
 206                 return -ENOMEM;
 207
 208         ref->root_id = root_id;
 209         if (key) {
 210                 ref->key_for_search = *key;
 211                 /*
 212                  * We can often find data backrefs with an offset that is too
 213                  * large (>= LLONG_MAX, maximum allowed file offset) due to
 214                  * underflows when subtracting a file's offset with the data
 215                  * offset of its corresponding extent data item. This can
 216                  * happen for example in the clone ioctl.
 217                  * So if we detect such case we set the search key's offset to
 218                  * zero to make sure we will find the matching file extent item
 219                  * at add_all_parents(), otherwise we will miss it because the
 220                  * offset taken form the backref is much larger then the offset
 221                  * of the file extent item. This can make us scan a very large
 222                  * number of file extent items, but at least it will not make
 223                  * us miss any.
 224                  * This is an ugly workaround for a behaviour that should have
 225                  * never existed, but it does and a fix for the clone ioctl
 226                  * would touch a lot of places, cause backwards incompatibility
 227                  * and would not fix the problem for extents cloned with older
 228                  * kernels.
 229                  */
 230                 if (ref->key_for_search.type == BTRFS_EXTENT_DATA_KEY &&
 231                     ref->key_for_search.offset >= LLONG_MAX)
 232                         ref->key_for_search.offset = 0;
 233         } else {
 234                 memset(&ref->key_for_search, 0, sizeof(ref->key_for_search));
 235         }
 236
 237         ref->inode_list = NULL;
 238         ref->level = level;
 239         ref->count = count;
 240         ref->parent = parent;
 241         ref->wanted_disk_byte = wanted_disk_byte;
 242         list_add_tail(&ref->list, head);
 243
 244         return 0;
 245 }
 246
 247 static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path,
 248                            struct ulist *parents, struct __prelim_ref *ref,
 249                            int level, u64 time_seq, const u64 *extent_item_pos,
 250                            u64 total_refs)
 251 {
 252         int ret = 0;
 253         int slot;
 254         struct extent_buffer *eb;
 255         struct btrfs_key key;
 256         struct btrfs_key *key_for_search = &ref->key_for_search;
 257         struct btrfs_file_extent_item *fi;
 258         struct extent_inode_elem *eie = NULL, *old = NULL;
 259         u64 disk_byte;
 260         u64 wanted_disk_byte = ref->wanted_disk_byte;
 261         u64 count = 0;
 262
 263         if (level != 0) {
 264                 eb = path->nodes[level];
 265                 ret = ulist_add(parents, eb->start, 0, GFP_NOFS);
 266                 if (ret < 0)
 267                         return ret;
 268                 return 0;
 269         }
 270
 271         /*
 272          * We normally enter this function with the path already pointing to
 273          * the first item to check. But sometimes, we may enter it with
 274          * slot==nritems. In that case, go to the next leaf before we continue.
 275          */
 276         if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) {
 277                 if (time_seq == (u64)-1)
 278                         ret = btrfs_next_leaf(root, path);
 279                 else
 280                         ret = btrfs_next_old_leaf(root, path, time_seq);
 281         }
 282
 283         while (!ret && count < total_refs) {
 284                 eb = path->nodes[0];
 285                 slot = path->slots[0];
 286
 287                 btrfs_item_key_to_cpu(eb, &key, slot);
 288
 289                 if (key.objectid != key_for_search->objectid ||
 290                     key.type != BTRFS_EXTENT_DATA_KEY)
 291                         break;
 292
 293                 fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
 294                 disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
 295
 296                 if (disk_byte == wanted_disk_byte) {
 297                         eie = NULL;
 298                         old = NULL;
 299                         count++;
 300                         if (extent_item_pos) {
 301                                 ret = check_extent_in_eb(&key, eb, fi,
 302                                                 *extent_item_pos,
 303                                                 &eie);
 304                                 if (ret < 0)
 305                                         break;
 306                         }
 307                         if (ret > 0)
 308                                 goto next;
 309                         ret = ulist_add_merge_ptr(parents, eb->start,
 310                                                   eie, (void **)&old, GFP_NOFS);
 311                         if (ret < 0)
 312                                 break;
 313                         if (!ret && extent_item_pos) {
 314                                 while (old->next)
 315                                         old = old->next;
 316                                 old->next = eie;
 317                         }
 318                         eie = NULL;
 319                 }
 320 next:
 321                 if (time_seq == (u64)-1)
 322                         ret = btrfs_next_item(root, path);
 323                 else
 324                         ret = btrfs_next_old_item(root, path, time_seq);
 325         }
 326
 327         if (ret > 0)
 328                 ret = 0;
 329         else if (ret < 0)
 330                 free_inode_elem_list(eie);
 331         return ret;
 332 }
 333
 334 /*
 335  * resolve an indirect backref in the form (root_id, key, level)
 336  * to a logical address
 337  */
 338 static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
 339                                   struct btrfs_path *path, u64 time_seq,
 340                                   struct __prelim_ref *ref,
 341                                   struct ulist *parents,
 342                                   const u64 *extent_item_pos, u64 total_refs)
 343 {
 344         struct btrfs_root *root;
 345         struct btrfs_key root_key;
 346         struct extent_buffer *eb;
 347         int ret = 0;
 348         int root_level;
 349         int level = ref->level;
 350         int index;
 351
 352         root_key.objectid = ref->root_id;
 353         root_key.type = BTRFS_ROOT_ITEM_KEY;
 354         root_key.offset = (u64)-1;
 355
 356         index = srcu_read_lock(&fs_info->subvol_srcu);
 357
 358         root = btrfs_read_fs_root_no_name(fs_info, &root_key);
 359         if (IS_ERR(root)) {
 360                 srcu_read_unlock(&fs_info->subvol_srcu, index);
 361                 ret = PTR_ERR(root);
 362                 goto out;
 363         }
 364
 365         if (path->search_commit_root)
 366                 root_level = btrfs_header_level(root->commit_root);
 367         else if (time_seq == (u64)-1)
 368                 root_level = btrfs_header_level(root->node);
 369         else
 370                 root_level = btrfs_old_root_level(root, time_seq);
 371
 372         if (root_level + 1 == level) {
 373                 srcu_read_unlock(&fs_info->subvol_srcu, index);
 374                 goto out;
 375         }
 376
 377         path->lowest_level = level;
 378         if (time_seq == (u64)-1)
 379                 ret = btrfs_search_slot(NULL, root, &ref->key_for_search, path,
 380                                         0, 0);
 381         else
 382                 ret = btrfs_search_old_slot(root, &ref->key_for_search, path,
 383                                             time_seq);
 384
 385         /* root node has been locked, we can release @subvol_srcu safely here */
 386         srcu_read_unlock(&fs_info->subvol_srcu, index);
 387
 388         pr_debug("search slot in root %llu (level %d, ref count %d) returned "
 389                  "%d for key (%llu %u %llu)\n",
 390                  ref->root_id, level, ref->count, ret,
 391                  ref->key_for_search.objectid, ref->key_for_search.type,
 392                  ref->key_for_search.offset);
 393         if (ret < 0)
 394                 goto out;
 395
 396         eb = path->nodes[level];
 397         while (!eb) {
 398                 if (WARN_ON(!level)) {
 399                         ret = 1;
 400                         goto out;
 401                 }
 402                 level--;
 403                 eb = path->nodes[level];
 404         }
 405
 406         ret = add_all_parents(root, path, parents, ref, level, time_seq,
 407                               extent_item_pos, total_refs);
 408 out:
 409         path->lowest_level = 0;
 410         btrfs_release_path(path);
 411         return ret;
 412 }
 413
 414 /*
 415  * resolve all indirect backrefs from the list
 416  */
 417 static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info,
 418                                    struct btrfs_path *path, u64 time_seq,
 419                                    struct list_head *head,
 420                                    const u64 *extent_item_pos, u64 total_refs,
 421                                    u64 root_objectid)
 422 {
 423         int err;
 424         int ret = 0;
 425         struct __prelim_ref *ref;
 426         struct __prelim_ref *ref_safe;
 427         struct __prelim_ref *new_ref;
 428         struct ulist *parents;
 429         struct ulist_node *node;
 430         struct ulist_iterator uiter;
 431
 432         parents = ulist_alloc(GFP_NOFS);
 433         if (!parents)
 434                 return -ENOMEM;
 435
 436         /*
 437          * _safe allows us to insert directly after the current item without
 438          * iterating over the newly inserted items.
 439          * we're also allowed to re-assign ref during iteration.
 440          */
 441         list_for_each_entry_safe(ref, ref_safe, head, list) {
 442                 if (ref->parent)        /* already direct */
 443                         continue;
 444                 if (ref->count == 0)
 445                         continue;
 446                 if (root_objectid && ref->root_id != root_objectid) {
 447                         ret = BACKREF_FOUND_SHARED;
 448                         goto out;
 449                 }
 450                 err = __resolve_indirect_ref(fs_info, path, time_seq, ref,
 451                                              parents, extent_item_pos,
 452                                              total_refs);
 453                 /*
 454                  * we can only tolerate ENOENT,otherwise,we should catch error
 455                  * and return directly.
 456                  */
 457                 if (err == -ENOENT) {
 458                         continue;
 459                 } else if (err) {
 460                         ret = err;
 461                         goto out;
 462                 }
 463
 464                 /* we put the first parent into the ref at hand */
 465                 ULIST_ITER_INIT(&uiter);
 466                 node = ulist_next(parents, &uiter);
 467                 ref->parent = node ? node->val : 0;
 468                 ref->inode_list = node ?
 469                         (struct extent_inode_elem *)(uintptr_t)node->aux : NULL;
 470
 471                 /* additional parents require new refs being added here */
 472                 while ((node = ulist_next(parents, &uiter))) {
 473                         new_ref = kmem_cache_alloc(btrfs_prelim_ref_cache,
 474                                                    GFP_NOFS);
 475                         if (!new_ref) {
 476                                 ret = -ENOMEM;
 477                                 goto out;
 478                         }
 479                         memcpy(new_ref, ref, sizeof(*ref));
 480                         new_ref->parent = node->val;
 481                         new_ref->inode_list = (struct extent_inode_elem *)
 482                                                         (uintptr_t)node->aux;
 483                         list_add(&new_ref->list, &ref->list);
 484                 }
 485                 ulist_reinit(parents);
 486         }
 487 out:
 488         ulist_free(parents);
 489         return ret;
 490 }
 491
 492 static inline int ref_for_same_block(struct __prelim_ref *ref1,
 493                                      struct __prelim_ref *ref2)
 494 {
 495         if (ref1->level != ref2->level)
 496                 return 0;
 497         if (ref1->root_id != ref2->root_id)
 498                 return 0;
 499         if (ref1->key_for_search.type != ref2->key_for_search.type)
 500                 return 0;
 501         if (ref1->key_for_search.objectid != ref2->key_for_search.objectid)
 502                 return 0;
 503         if (ref1->key_for_search.offset != ref2->key_for_search.offset)
 504                 return 0;
 505         if (ref1->parent != ref2->parent)
 506                 return 0;
 507
 508         return 1;
 509 }
 510
 511 /*
 512  * read tree blocks and add keys where required.
 513  */
 514 static int __add_missing_keys(struct btrfs_fs_info *fs_info,
 515                               struct list_head *head)
 516 {
 517         struct list_head *pos;
 518         struct extent_buffer *eb;
 519
 520         list_for_each(pos, head) {
 521                 struct __prelim_ref *ref;
 522                 ref = list_entry(pos, struct __prelim_ref, list);
 523
 524                 if (ref->parent)
 525                         continue;
 526                 if (ref->key_for_search.type)
 527                         continue;
 528                 BUG_ON(!ref->wanted_disk_byte);
 529                 eb = read_tree_block(fs_info->tree_root, ref->wanted_disk_byte,
 530                                      0);
 531                 if (IS_ERR(eb)) {
 532                         return PTR_ERR(eb);
 533                 } else if (!extent_buffer_uptodate(eb)) {
 534                         free_extent_buffer(eb);
 535                         return -EIO;
 536                 }
 537                 btrfs_tree_read_lock(eb);
 538                 if (btrfs_header_level(eb) == 0)
 539                         btrfs_item_key_to_cpu(eb, &ref->key_for_search, 0);
 540                 else
 541                         btrfs_node_key_to_cpu(eb, &ref->key_for_search, 0);
 542                 btrfs_tree_read_unlock(eb);
 543                 free_extent_buffer(eb);
 544         }
 545         return 0;
 546 }
 547
 548 /*
 549  * merge backrefs and adjust counts accordingly
 550  *
 551  * mode = 1: merge identical keys, if key is set
 552  *    FIXME: if we add more keys in __add_prelim_ref, we can merge more here.
 553  *           additionally, we could even add a key range for the blocks we
 554  *           looked into to merge even more (-> replace unresolved refs by those
 555  *           having a parent).
 556  * mode = 2: merge identical parents
 557  */
 558 static void __merge_refs(struct list_head *head, int mode)
 559 {
 560         struct list_head *pos1;
 561
 562         list_for_each(pos1, head) {
 563                 struct list_head *n2;
 564                 struct list_head *pos2;
 565                 struct __prelim_ref *ref1;
 566
 567                 ref1 = list_entry(pos1, struct __prelim_ref, list);
 568
 569                 for (pos2 = pos1->next, n2 = pos2->next; pos2 != head;
 570                      pos2 = n2, n2 = pos2->next) {
 571                         struct __prelim_ref *ref2;
 572                         struct __prelim_ref *xchg;
 573                         struct extent_inode_elem *eie;
 574
 575                         ref2 = list_entry(pos2, struct __prelim_ref, list);
 576
 577                         if (!ref_for_same_block(ref1, ref2))
 578                                 continue;
 579                         if (mode == 1) {
 580                                 if (!ref1->parent && ref2->parent) {
 581                                         xchg = ref1;
 582                                         ref1 = ref2;
 583                                         ref2 = xchg;
 584                                 }
 585                         } else {
 586                                 if (ref1->parent != ref2->parent)
 587                                         continue;
 588                         }
 589
 590                         eie = ref1->inode_list;
 591                         while (eie && eie->next)
 592                                 eie = eie->next;
 593                         if (eie)
 594                                 eie->next = ref2->inode_list;
 595                         else
 596                                 ref1->inode_list = ref2->inode_list;
 597                         ref1->count += ref2->count;
 598
 599                         list_del(&ref2->list);
 600                         kmem_cache_free(btrfs_prelim_ref_cache, ref2);
 601                 }
 602
 603         }
 604 }
 605
 606 /*
 607  * add all currently queued delayed refs from this head whose seq nr is
 608  * smaller or equal that seq to the list
 609  */
 610 static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
 611                               struct list_head *prefs, u64 *total_refs,
 612                               u64 inum)
 613 {
 614         struct btrfs_delayed_ref_node *node;
 615         struct btrfs_delayed_extent_op *extent_op = head->extent_op;
 616         struct btrfs_key key;
 617         struct btrfs_key op_key = {0};
 618         int sgn;
 619         int ret = 0;
 620
 621         if (extent_op && extent_op->update_key)
 622                 btrfs_disk_key_to_cpu(&op_key, &extent_op->key);
 623
 624         spin_lock(&head->lock);
 625         list_for_each_entry(node, &head->ref_list, list) {
 626                 if (node->seq > seq)
 627                         continue;
 628
 629                 switch (node->action) {
 630                 case BTRFS_ADD_DELAYED_EXTENT:
 631                 case BTRFS_UPDATE_DELAYED_HEAD:
 632                         WARN_ON(1);
 633                         continue;
 634                 case BTRFS_ADD_DELAYED_REF:
 635                         sgn = 1;
 636                         break;
 637                 case BTRFS_DROP_DELAYED_REF:
 638                         sgn = -1;
 639                         break;
 640                 default:
 641                         BUG_ON(1);
 642                 }
 643                 *total_refs += (node->ref_mod * sgn);
 644                 switch (node->type) {
 645                 case BTRFS_TREE_BLOCK_REF_KEY: {
 646                         struct btrfs_delayed_tree_ref *ref;
 647
 648                         ref = btrfs_delayed_node_to_tree_ref(node);
 649                         ret = __add_prelim_ref(prefs, ref->root, &op_key,
 650                                                ref->level + 1, 0, node->bytenr,
 651                                                node->ref_mod * sgn, GFP_ATOMIC);
 652                         break;
 653                 }
 654                 case BTRFS_SHARED_BLOCK_REF_KEY: {
 655                         struct btrfs_delayed_tree_ref *ref;
 656
 657                         ref = btrfs_delayed_node_to_tree_ref(node);
 658                         ret = __add_prelim_ref(prefs, 0, NULL,
 659                                                ref->level + 1, ref->parent,
 660                                                node->bytenr,
 661                                                node->ref_mod * sgn, GFP_ATOMIC);
 662                         break;
 663                 }
 664                 case BTRFS_EXTENT_DATA_REF_KEY: {
 665                         struct btrfs_delayed_data_ref *ref;
 666                         ref = btrfs_delayed_node_to_data_ref(node);
 667
 668                         key.objectid = ref->objectid;
 669                         key.type = BTRFS_EXTENT_DATA_KEY;
 670                         key.offset = ref->offset;
 671
 672                         /*
 673                          * Found a inum that doesn't match our known inum, we
 674                          * know it's shared.
 675                          */
 676                         if (inum && ref->objectid != inum) {
 677                                 ret = BACKREF_FOUND_SHARED;
 678                                 break;
 679                         }
 680
 681                         ret = __add_prelim_ref(prefs, ref->root, &key, 0, 0,
 682                                                node->bytenr,
 683                                                node->ref_mod * sgn, GFP_ATOMIC);
 684                         break;
 685                 }
 686                 case BTRFS_SHARED_DATA_REF_KEY: {
 687                         struct btrfs_delayed_data_ref *ref;
 688
 689                         ref = btrfs_delayed_node_to_data_ref(node);
 690                         ret = __add_prelim_ref(prefs, 0, NULL, 0,
 691                                                ref->parent, node->bytenr,
 692                                                node->ref_mod * sgn, GFP_ATOMIC);
 693                         break;
 694                 }
 695                 default:
 696                         WARN_ON(1);
 697                 }
 698                 if (ret)
 699                         break;
 700         }
 701         spin_unlock(&head->lock);
 702         return ret;
 703 }
 704
 705 /*
 706  * add all inline backrefs for bytenr to the list
 707  */
 708 static int __add_inline_refs(struct btrfs_fs_info *fs_info,
 709                              struct btrfs_path *path, u64 bytenr,
 710                              int *info_level, struct list_head *prefs,
 711                              u64 *total_refs, u64 inum)
 712 {
 713         int ret = 0;
 714         int slot;
 715         struct extent_buffer *leaf;
 716         struct btrfs_key key;
 717         struct btrfs_key found_key;
 718         unsigned long ptr;
 719         unsigned long end;
 720         struct btrfs_extent_item *ei;
 721         u64 flags;
 722         u64 item_size;
 723
 724         /*
 725          * enumerate all inline refs
 726          */
 727         leaf = path->nodes[0];
 728         slot = path->slots[0];
 729
 730         item_size = btrfs_item_size_nr(leaf, slot);
 731         BUG_ON(item_size < sizeof(*ei));
 732
 733         ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item);
 734         flags = btrfs_extent_flags(leaf, ei);
 735         *total_refs += btrfs_extent_refs(leaf, ei);
 736         btrfs_item_key_to_cpu(leaf, &found_key, slot);
 737
 738         ptr = (unsigned long)(ei + 1);
 739         end = (unsigned long)ei + item_size;
 740
 741         if (found_key.type == BTRFS_EXTENT_ITEM_KEY &&
 742             flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
 743                 struct btrfs_tree_block_info *info;
 744
 745                 info = (struct btrfs_tree_block_info *)ptr;
 746                 *info_level = btrfs_tree_block_level(leaf, info);
 747                 ptr += sizeof(struct btrfs_tree_block_info);
 748                 BUG_ON(ptr > end);
 749         } else if (found_key.type == BTRFS_METADATA_ITEM_KEY) {
 750                 *info_level = found_key.offset;
 751         } else {
 752                 BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA));
 753         }
 754
 755         while (ptr < end) {
 756                 struct btrfs_extent_inline_ref *iref;
 757                 u64 offset;
 758                 int type;
 759
 760                 iref = (struct btrfs_extent_inline_ref *)ptr;
 761                 type = btrfs_extent_inline_ref_type(leaf, iref);
 762                 offset = btrfs_extent_inline_ref_offset(leaf, iref);
 763
 764                 switch (type) {
 765                 case BTRFS_SHARED_BLOCK_REF_KEY:
 766                         ret = __add_prelim_ref(prefs, 0, NULL,
 767                                                 *info_level + 1, offset,
 768                                                 bytenr, 1, GFP_NOFS);
 769                         break;
 770                 case BTRFS_SHARED_DATA_REF_KEY: {
 771                         struct btrfs_shared_data_ref *sdref;
 772                         int count;
 773
 774                         sdref = (struct btrfs_shared_data_ref *)(iref + 1);
 775                         count = btrfs_shared_data_ref_count(leaf, sdref);
 776                         ret = __add_prelim_ref(prefs, 0, NULL, 0, offset,
 777                                                bytenr, count, GFP_NOFS);
 778                         break;
 779                 }
 780                 case BTRFS_TREE_BLOCK_REF_KEY:
 781                         ret = __add_prelim_ref(prefs, offset, NULL,
 782                                                *info_level + 1, 0,
 783                                                bytenr, 1, GFP_NOFS);
 784                         break;
 785                 case BTRFS_EXTENT_DATA_REF_KEY: {
 786                         struct btrfs_extent_data_ref *dref;
 787                         int count;
 788                         u64 root;
 789
 790                         dref = (struct btrfs_extent_data_ref *)(&iref->offset);
 791                         count = btrfs_extent_data_ref_count(leaf, dref);
 792                         key.objectid = btrfs_extent_data_ref_objectid(leaf,
 793                                                                       dref);
 794                         key.type = BTRFS_EXTENT_DATA_KEY;
 795                         key.offset = btrfs_extent_data_ref_offset(leaf, dref);
 796
 797                         if (inum && key.objectid != inum) {
 798                                 ret = BACKREF_FOUND_SHARED;
 799                                 break;
 800                         }
 801
 802                         root = btrfs_extent_data_ref_root(leaf, dref);
 803                         ret = __add_prelim_ref(prefs, root, &key, 0, 0,
 804                                                bytenr, count, GFP_NOFS);
 805                         break;
 806                 }
 807                 default:
 808                         WARN_ON(1);
 809                 }
 810                 if (ret)
 811                         return ret;
 812                 ptr += btrfs_extent_inline_ref_size(type);
 813         }
 814
 815         return 0;
 816 }
 817
 818 /*
 819  * add all non-inline backrefs for bytenr to the list
 820  */
 821 static int __add_keyed_refs(struct btrfs_fs_info *fs_info,
 822                             struct btrfs_path *path, u64 bytenr,
 823                             int info_level, struct list_head *prefs, u64 inum)
 824 {
 825         struct btrfs_root *extent_root = fs_info->extent_root;
 826         int ret;
 827         int slot;
 828         struct extent_buffer *leaf;
 829         struct btrfs_key key;
 830
 831         while (1) {
 832                 ret = btrfs_next_item(extent_root, path);
 833                 if (ret < 0)
 834                         break;
 835                 if (ret) {
 836                         ret = 0;
 837                         break;
 838                 }
 839
 840                 slot = path->slots[0];
 841                 leaf = path->nodes[0];
 842                 btrfs_item_key_to_cpu(leaf, &key, slot);
 843
 844                 if (key.objectid != bytenr)
 845                         break;
 846                 if (key.type < BTRFS_TREE_BLOCK_REF_KEY)
 847                         continue;
 848                 if (key.type > BTRFS_SHARED_DATA_REF_KEY)
 849                         break;
 850
 851                 switch (key.type) {
 852                 case BTRFS_SHARED_BLOCK_REF_KEY:
 853                         ret = __add_prelim_ref(prefs, 0, NULL,
 854                                                 info_level + 1, key.offset,
 855                                                 bytenr, 1, GFP_NOFS);
 856                         break;
 857                 case BTRFS_SHARED_DATA_REF_KEY: {
 858                         struct btrfs_shared_data_ref *sdref;
 859                         int count;
 860
 861                         sdref = btrfs_item_ptr(leaf, slot,
 862                                               struct btrfs_shared_data_ref);
 863                         count = btrfs_shared_data_ref_count(leaf, sdref);
 864                         ret = __add_prelim_ref(prefs, 0, NULL, 0, key.offset,
 865                                                 bytenr, count, GFP_NOFS);
 866                         break;
 867                 }
 868                 case BTRFS_TREE_BLOCK_REF_KEY:
 869                         ret = __add_prelim_ref(prefs, key.offset, NULL,
 870                                                info_level + 1, 0,
 871                                                bytenr, 1, GFP_NOFS);
 872                         break;
 873                 case BTRFS_EXTENT_DATA_REF_KEY: {
 874                         struct btrfs_extent_data_ref *dref;
 875                         int count;
 876                         u64 root;
 877
 878                         dref = btrfs_item_ptr(leaf, slot,
 879                                               struct btrfs_extent_data_ref);
 880                         count = btrfs_extent_data_ref_count(leaf, dref);
 881                         key.objectid = btrfs_extent_data_ref_objectid(leaf,
 882                                                                       dref);
 883                         key.type = BTRFS_EXTENT_DATA_KEY;
 884                         key.offset = btrfs_extent_data_ref_offset(leaf, dref);
 885
 886                         if (inum && key.objectid != inum) {
 887                                 ret = BACKREF_FOUND_SHARED;
 888                                 break;
 889                         }
 890
 891                         root = btrfs_extent_data_ref_root(leaf, dref);
 892                         ret = __add_prelim_ref(prefs, root, &key, 0, 0,
 893                                                bytenr, count, GFP_NOFS);
 894                         break;
 895                 }
 896                 default:
 897                         WARN_ON(1);
 898                 }
 899                 if (ret)
 900                         return ret;
 901
 902         }
 903
 904         return ret;
 905 }
 906
 907 /*
 908  * this adds all existing backrefs (inline backrefs, backrefs and delayed
 909  * refs) for the given bytenr to the refs list, merges duplicates and resolves
 910  * indirect refs to their parent bytenr.
 911  * When roots are found, they're added to the roots list
 912  *
 913  * NOTE: This can return values > 0
 914  *
 915  * If time_seq is set to (u64)-1, it will not search delayed_refs, and behave
 916  * much like trans == NULL case, the difference only lies in it will not
 917  * commit root.
 918  * The special case is for qgroup to search roots in commit_transaction().
 919  *
 920  * FIXME some caching might speed things up
 921  */
 922 static int find_parent_nodes(struct btrfs_trans_handle *trans,
 923                              struct btrfs_fs_info *fs_info, u64 bytenr,
 924                              u64 time_seq, struct ulist *refs,
 925                              struct ulist *roots, const u64 *extent_item_pos,
 926                              u64 root_objectid, u64 inum)
 927 {
 928         struct btrfs_key key;
 929         struct btrfs_path *path;
 930         struct btrfs_delayed_ref_root *delayed_refs = NULL;
 931         struct btrfs_delayed_ref_head *head;
 932         int info_level = 0;
 933         int ret;
 934         struct list_head prefs_delayed;
 935         struct list_head prefs;
 936         struct __prelim_ref *ref;
 937         struct extent_inode_elem *eie = NULL;
 938         u64 total_refs = 0;
 939
 940         INIT_LIST_HEAD(&prefs);
 941         INIT_LIST_HEAD(&prefs_delayed);
 942
 943         key.objectid = bytenr;
 944         key.offset = (u64)-1;
 945         if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
 946                 key.type = BTRFS_METADATA_ITEM_KEY;
 947         else
 948                 key.type = BTRFS_EXTENT_ITEM_KEY;
 949
 950         path = btrfs_alloc_path();
 951         if (!path)
 952                 return -ENOMEM;
 953         if (!trans) {
 954                 path->search_commit_root = 1;
 955                 path->skip_locking = 1;
 956         }
 957
 958         if (time_seq == (u64)-1)
 959                 path->skip_locking = 1;
 960
 961         /*
 962          * grab both a lock on the path and a lock on the delayed ref head.
 963          * We need both to get a consistent picture of how the refs look
 964          * at a specified point in time
 965          */
 966 again:
 967         head = NULL;
 968
 969         ret = btrfs_search_slot(trans, fs_info->extent_root, &key, path, 0, 0);
 970         if (ret < 0)
 971                 goto out;
 972         BUG_ON(ret == 0);
 973
 974 #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
 975         if (trans && likely(trans->type != __TRANS_DUMMY) &&
 976             time_seq != (u64)-1) {
 977 #else
 978         if (trans && time_seq != (u64)-1) {
 979 #endif
 980                 /*
 981                  * look if there are updates for this ref queued and lock the
 982                  * head
 983                  */
 984                 delayed_refs = &trans->transaction->delayed_refs;
 985                 spin_lock(&delayed_refs->lock);
 986                 head = btrfs_find_delayed_ref_head(trans, bytenr);
 987                 if (head) {
 988                         if (!mutex_trylock(&head->mutex)) {
 989                                 atomic_inc(&head->node.refs);
 990                                 spin_unlock(&delayed_refs->lock);
 991
 992                                 btrfs_release_path(path);
 993
 994                                 /*
 995                                  * Mutex was contended, block until it's
 996                                  * released and try again
 997                                  */
 998                                 mutex_lock(&head->mutex);
 999                                 mutex_unlock(&head->mutex);
1000                                 btrfs_put_delayed_ref(&head->node);
1001                                 goto again;
1002                         }
1003                         spin_unlock(&delayed_refs->lock);
1004                         ret = __add_delayed_refs(head, time_seq,
1005                                                  &prefs_delayed, &total_refs,
1006                                                  inum);
1007                         mutex_unlock(&head->mutex);
1008                         if (ret)
1009                                 goto out;
1010                 } else {
1011                         spin_unlock(&delayed_refs->lock);
1012                 }
1013         }
1014
1015         if (path->slots[0]) {
1016                 struct extent_buffer *leaf;
1017                 int slot;
1018
1019                 path->slots[0]--;
1020                 leaf = path->nodes[0];
1021                 slot = path->slots[0];
1022                 btrfs_item_key_to_cpu(leaf, &key, slot);
1023                 if (key.objectid == bytenr &&
1024                     (key.type == BTRFS_EXTENT_ITEM_KEY ||
1025                      key.type == BTRFS_METADATA_ITEM_KEY)) {
1026                         ret = __add_inline_refs(fs_info, path, bytenr,
1027                                                 &info_level, &prefs,
1028                                                 &total_refs, inum);
1029                         if (ret)
1030                                 goto out;
1031                         ret = __add_keyed_refs(fs_info, path, bytenr,
1032                                                info_level, &prefs, inum);
1033                         if (ret)
1034                                 goto out;
1035                 }
1036         }
1037         btrfs_release_path(path);
1038
1039         list_splice_init(&prefs_delayed, &prefs);
1040
1041         ret = __add_missing_keys(fs_info, &prefs);
1042         if (ret)
1043                 goto out;
1044
1045         __merge_refs(&prefs, 1);
1046
1047         ret = __resolve_indirect_refs(fs_info, path, time_seq, &prefs,
1048                                       extent_item_pos, total_refs,
1049                                       root_objectid);
1050         if (ret)
1051                 goto out;
1052
1053         __merge_refs(&prefs, 2);
1054
1055         while (!list_empty(&prefs)) {
1056                 ref = list_first_entry(&prefs, struct __prelim_ref, list);
1057                 WARN_ON(ref->count < 0);
1058                 if (roots && ref->count && ref->root_id && ref->parent == 0) {
1059                         if (root_objectid && ref->root_id != root_objectid) {
1060                                 ret = BACKREF_FOUND_SHARED;
1061                                 goto out;
1062                         }
1063
1064                         /* no parent == root of tree */
1065                         ret = ulist_add(roots, ref->root_id, 0, GFP_NOFS);
1066                         if (ret < 0)
1067                                 goto out;
1068                 }
1069                 if (ref->count && ref->parent) {
1070                         if (extent_item_pos && !ref->inode_list &&
1071                             ref->level == 0) {
1072                                 struct extent_buffer *eb;
1073
1074                                 eb = read_tree_block(fs_info->extent_root,
1075                                                            ref->parent, 0);
1076                                 if (IS_ERR(eb)) {
1077                                         ret = PTR_ERR(eb);
1078                                         goto out;
1079                                 } else if (!extent_buffer_uptodate(eb)) {
1080                                         free_extent_buffer(eb);
1081                                         ret = -EIO;
1082                                         goto out;
1083                                 }
1084                                 btrfs_tree_read_lock(eb);
1085                                 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1086                                 ret = find_extent_in_eb(eb, bytenr,
1087                                                         *extent_item_pos, &eie);
1088                                 btrfs_tree_read_unlock_blocking(eb);
1089                                 free_extent_buffer(eb);
1090                                 if (ret < 0)
1091                                         goto out;
1092                                 ref->inode_list = eie;
1093                         }
1094                         ret = ulist_add_merge_ptr(refs, ref->parent,
1095                                                   ref->inode_list,
1096                                                   (void **)&eie, GFP_NOFS);
1097                         if (ret < 0)
1098                                 goto out;
1099                         if (!ret && extent_item_pos) {
1100                                 /*
1101                                  * we've recorded that parent, so we must extend
1102                                  * its inode list here
1103                                  */
1104                                 BUG_ON(!eie);
1105                                 while (eie->next)
1106                                         eie = eie->next;
1107                                 eie->next = ref->inode_list;
1108                         }
1109                         eie = NULL;
1110                 }
1111                 list_del(&ref->list);
1112                 kmem_cache_free(btrfs_prelim_ref_cache, ref);
1113         }
1114
1115 out:
1116         btrfs_free_path(path);
1117         while (!list_empty(&prefs)) {
1118                 ref = list_first_entry(&prefs, struct __prelim_ref, list);
1119                 list_del(&ref->list);
1120                 kmem_cache_free(btrfs_prelim_ref_cache, ref);
1121         }
1122         while (!list_empty(&prefs_delayed)) {
1123                 ref = list_first_entry(&prefs_delayed, struct __prelim_ref,
1124                                        list);
1125                 list_del(&ref->list);
1126                 kmem_cache_free(btrfs_prelim_ref_cache, ref);
1127         }
1128         if (ret < 0)
1129                 free_inode_elem_list(eie);
1130         return ret;
1131 }
1132
1133 static void free_leaf_list(struct ulist *blocks)
1134 {
1135         struct ulist_node *node = NULL;
1136         struct extent_inode_elem *eie;
1137         struct ulist_iterator uiter;
1138
1139         ULIST_ITER_INIT(&uiter);
1140         while ((node = ulist_next(blocks, &uiter))) {
1141                 if (!node->aux)
1142                         continue;
1143                 eie = (struct extent_inode_elem *)(uintptr_t)node->aux;
1144                 free_inode_elem_list(eie);
1145                 node->aux = 0;
1146         }
1147
1148         ulist_free(blocks);
1149 }
1150
1151 /*
1152  * Finds all leafs with a reference to the specified combination of bytenr and
1153  * offset. key_list_head will point to a list of corresponding keys (caller must
1154  * free each list element). The leafs will be stored in the leafs ulist, which
1155  * must be freed with ulist_free.
1156  *
1157  * returns 0 on success, <0 on error
1158  */
1159 static int btrfs_find_all_leafs(struct btrfs_trans_handle *trans,
1160                                 struct btrfs_fs_info *fs_info, u64 bytenr,
1161                                 u64 time_seq, struct ulist **leafs,
1162                                 const u64 *extent_item_pos)
1163 {
1164         int ret;
1165
1166         *leafs = ulist_alloc(GFP_NOFS);
1167         if (!*leafs)
1168                 return -ENOMEM;
1169
1170         ret = find_parent_nodes(trans, fs_info, bytenr,
1171                                 time_seq, *leafs, NULL, extent_item_pos, 0, 0);
1172         if (ret < 0 && ret != -ENOENT) {
1173                 free_leaf_list(*leafs);
1174                 return ret;
1175         }
1176
1177         return 0;
1178 }
1179
1180 /*
1181  * walk all backrefs for a given extent to find all roots that reference this
1182  * extent. Walking a backref means finding all extents that reference this
1183  * extent and in turn walk the backrefs of those, too. Naturally this is a
1184  * recursive process, but here it is implemented in an iterative fashion: We
1185  * find all referencing extents for the extent in question and put them on a
1186  * list. In turn, we find all referencing extents for those, further appending
1187  * to the list. The way we iterate the list allows adding more elements after
1188  * the current while iterating. The process stops when we reach the end of the
1189  * list. Found roots are added to the roots list.
1190  *
1191  * returns 0 on success, < 0 on error.
1192  */
1193 static int __btrfs_find_all_roots(struct btrfs_trans_handle *trans,
1194                                   struct btrfs_fs_info *fs_info, u64 bytenr,
1195                                   u64 time_seq, struct ulist **roots)
1196 {
1197         struct ulist *tmp;
1198         struct ulist_node *node = NULL;
1199         struct ulist_iterator uiter;
1200         int ret;
1201
1202         tmp = ulist_alloc(GFP_NOFS);
1203         if (!tmp)
1204                 return -ENOMEM;
1205         *roots = ulist_alloc(GFP_NOFS);
1206         if (!*roots) {
1207                 ulist_free(tmp);
1208                 return -ENOMEM;
1209         }
1210
1211         ULIST_ITER_INIT(&uiter);
1212         while (1) {
1213                 ret = find_parent_nodes(trans, fs_info, bytenr,
1214                                         time_seq, tmp, *roots, NULL, 0, 0);
1215                 if (ret < 0 && ret != -ENOENT) {
1216                         ulist_free(tmp);
1217                         ulist_free(*roots);
1218                         return ret;
1219                 }
1220                 node = ulist_next(tmp, &uiter);
1221                 if (!node)
1222                         break;
1223                 bytenr = node->val;
1224                 cond_resched();
1225         }
1226
1227         ulist_free(tmp);
1228         return 0;
1229 }
1230
1231 int btrfs_find_all_roots(struct btrfs_trans_handle *trans,
1232                          struct btrfs_fs_info *fs_info, u64 bytenr,
1233                          u64 time_seq, struct ulist **roots)
1234 {
1235         int ret;
1236
1237         if (!trans)
1238                 down_read(&fs_info->commit_root_sem);
1239         ret = __btrfs_find_all_roots(trans, fs_info, bytenr, time_seq, roots);
1240         if (!trans)
1241                 up_read(&fs_info->commit_root_sem);
1242         return ret;
1243 }
1244
1245 /**
1246  * btrfs_check_shared - tell us whether an extent is shared
1247  *
1248  * @trans: optional trans handle
1249  *
1250  * btrfs_check_shared uses the backref walking code but will short
1251  * circuit as soon as it finds a root or inode that doesn't match the
1252  * one passed in. This provides a significant performance benefit for
1253  * callers (such as fiemap) which want to know whether the extent is
1254  * shared but do not need a ref count.
1255  *
1256  * Return: 0 if extent is not shared, 1 if it is shared, < 0 on error.
1257  */
1258 int btrfs_check_shared(struct btrfs_trans_handle *trans,
1259                        struct btrfs_fs_info *fs_info, u64 root_objectid,
1260                        u64 inum, u64 bytenr)
1261 {
1262         struct ulist *tmp = NULL;
1263         struct ulist *roots = NULL;
1264         struct ulist_iterator uiter;
1265         struct ulist_node *node;
1266         struct seq_list elem = SEQ_LIST_INIT(elem);
1267         int ret = 0;
1268
1269         tmp = ulist_alloc(GFP_NOFS);
1270         roots = ulist_alloc(GFP_NOFS);
1271         if (!tmp || !roots) {
1272                 ulist_free(tmp);
1273                 ulist_free(roots);
1274                 return -ENOMEM;
1275         }
1276
1277         if (trans)
1278                 btrfs_get_tree_mod_seq(fs_info, &elem);
1279         else
1280                 down_read(&fs_info->commit_root_sem);
1281         ULIST_ITER_INIT(&uiter);
1282         while (1) {
1283                 ret = find_parent_nodes(trans, fs_info, bytenr, elem.seq, tmp,
1284                                         roots, NULL, root_objectid, inum);
1285                 if (ret == BACKREF_FOUND_SHARED) {
1286                         /* this is the only condition under which we return 1 */
1287                         ret = 1;
1288                         break;
1289                 }
1290                 if (ret < 0 && ret != -ENOENT)
1291                         break;
1292                 ret = 0;
1293                 node = ulist_next(tmp, &uiter);
1294                 if (!node)
1295                         break;
1296                 bytenr = node->val;
1297                 cond_resched();
1298         }
1299         if (trans)
1300                 btrfs_put_tree_mod_seq(fs_info, &elem);
1301         else
1302                 up_read(&fs_info->commit_root_sem);
1303         ulist_free(tmp);
1304         ulist_free(roots);
1305         return ret;
1306 }
1307
1308 int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid,
1309                           u64 start_off, struct btrfs_path *path,
1310                           struct btrfs_inode_extref **ret_extref,
1311                           u64 *found_off)
1312 {
1313         int ret, slot;
1314         struct btrfs_key key;
1315         struct btrfs_key found_key;
1316         struct btrfs_inode_extref *extref;
1317         struct extent_buffer *leaf;
1318         unsigned long ptr;
1319
1320         key.objectid = inode_objectid;
1321         key.type = BTRFS_INODE_EXTREF_KEY;
1322         key.offset = start_off;
1323
1324         ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1325         if (ret < 0)
1326                 return ret;
1327
1328         while (1) {
1329                 leaf = path->nodes[0];
1330                 slot = path->slots[0];
1331                 if (slot >= btrfs_header_nritems(leaf)) {
1332                         /*
1333                          * If the item at offset is not found,
1334                          * btrfs_search_slot will point us to the slot
1335                          * where it should be inserted. In our case
1336                          * that will be the slot directly before the
1337                          * next INODE_REF_KEY_V2 item. In the case
1338                          * that we're pointing to the last slot in a
1339                          * leaf, we must move one leaf over.
1340                          */
1341                         ret = btrfs_next_leaf(root, path);
1342                         if (ret) {
1343                                 if (ret >= 1)
1344                                         ret = -ENOENT;
1345                                 break;
1346                         }
1347                         continue;
1348                 }
1349
1350                 btrfs_item_key_to_cpu(leaf, &found_key, slot);
1351
1352                 /*
1353                  * Check that we're still looking at an extended ref key for
1354                  * this particular objectid. If we have different
1355                  * objectid or type then there are no more to be found
1356                  * in the tree and we can exit.
1357                  */
1358                 ret = -ENOENT;
1359                 if (found_key.objectid != inode_objectid)
1360                         break;
1361                 if (found_key.type != BTRFS_INODE_EXTREF_KEY)
1362                         break;
1363
1364                 ret = 0;
1365                 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
1366                 extref = (struct btrfs_inode_extref *)ptr;
1367                 *ret_extref = extref;
1368                 if (found_off)
1369                         *found_off = found_key.offset;
1370                 break;
1371         }
1372
1373         return ret;
1374 }
1375
1376 /*
1377  * this iterates to turn a name (from iref/extref) into a full filesystem path.
1378  * Elements of the path are separated by '/' and the path is guaranteed to be
1379  * 0-terminated. the path is only given within the current file system.
1380  * Therefore, it never starts with a '/'. the caller is responsible to provide
1381  * "size" bytes in "dest". the dest buffer will be filled backwards. finally,
1382  * the start point of the resulting string is returned. this pointer is within
1383  * dest, normally.
1384  * in case the path buffer would overflow, the pointer is decremented further
1385  * as if output was written to the buffer, though no more output is actually
1386  * generated. that way, the caller can determine how much space would be
1387  * required for the path to fit into the buffer. in that case, the returned
1388  * value will be smaller than dest. callers must check this!
1389  */
1390 char *btrfs_ref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
1391                         u32 name_len, unsigned long name_off,
1392                         struct extent_buffer *eb_in, u64 parent,
1393                         char *dest, u32 size)
1394 {
1395         int slot;
1396         u64 next_inum;
1397         int ret;
1398         s64 bytes_left = ((s64)size) - 1;
1399         struct extent_buffer *eb = eb_in;
1400         struct btrfs_key found_key;
1401         int leave_spinning = path->leave_spinning;
1402         struct btrfs_inode_ref *iref;
1403
1404         if (bytes_left >= 0)
1405                 dest[bytes_left] = '\0';
1406
1407         path->leave_spinning = 1;
1408         while (1) {
1409                 bytes_left -= name_len;
1410                 if (bytes_left >= 0)
1411                         read_extent_buffer(eb, dest + bytes_left,
1412                                            name_off, name_len);
1413                 if (eb != eb_in) {
1414                         btrfs_tree_read_unlock_blocking(eb);
1415                         free_extent_buffer(eb);
1416                 }
1417                 ret = btrfs_find_item(fs_root, path, parent, 0,
1418                                 BTRFS_INODE_REF_KEY, &found_key);
1419                 if (ret > 0)
1420                         ret = -ENOENT;
1421                 if (ret)
1422                         break;
1423
1424                 next_inum = found_key.offset;
1425
1426                 /* regular exit ahead */
1427                 if (parent == next_inum)
1428                         break;
1429
1430                 slot = path->slots[0];
1431                 eb = path->nodes[0];
1432                 /* make sure we can use eb after releasing the path */
1433                 if (eb != eb_in) {
1434                         atomic_inc(&eb->refs);
1435                         btrfs_tree_read_lock(eb);
1436                         btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1437                 }
1438                 btrfs_release_path(path);
1439                 iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
1440
1441                 name_len = btrfs_inode_ref_name_len(eb, iref);
1442                 name_off = (unsigned long)(iref + 1);
1443
1444                 parent = next_inum;
1445                 --bytes_left;
1446                 if (bytes_left >= 0)
1447                         dest[bytes_left] = '/';
1448         }
1449
1450         btrfs_release_path(path);
1451         path->leave_spinning = leave_spinning;
1452
1453         if (ret)
1454                 return ERR_PTR(ret);
1455
1456         return dest + bytes_left;
1457 }
1458
1459 /*
1460  * this makes the path point to (logical EXTENT_ITEM *)
1461  * returns BTRFS_EXTENT_FLAG_DATA for data, BTRFS_EXTENT_FLAG_TREE_BLOCK for
1462  * tree blocks and <0 on error.
1463  */
1464 int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
1465                         struct btrfs_path *path, struct btrfs_key *found_key,
1466                         u64 *flags_ret)
1467 {
1468         int ret;
1469         u64 flags;
1470         u64 size = 0;
1471         u32 item_size;
1472         struct extent_buffer *eb;
1473         struct btrfs_extent_item *ei;
1474         struct btrfs_key key;
1475
1476         if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
1477                 key.type = BTRFS_METADATA_ITEM_KEY;
1478         else
1479                 key.type = BTRFS_EXTENT_ITEM_KEY;
1480         key.objectid = logical;
1481         key.offset = (u64)-1;
1482
1483         ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0);
1484         if (ret < 0)
1485                 return ret;
1486
1487         ret = btrfs_previous_extent_item(fs_info->extent_root, path, 0);
1488         if (ret) {
1489                 if (ret > 0)
1490                         ret = -ENOENT;
1491                 return ret;
1492         }
1493         btrfs_item_key_to_cpu(path->nodes[0], found_key, path->slots[0]);
1494         if (found_key->type == BTRFS_METADATA_ITEM_KEY)
1495                 size = fs_info->extent_root->nodesize;
1496         else if (found_key->type == BTRFS_EXTENT_ITEM_KEY)
1497                 size = found_key->offset;
1498
1499         if (found_key->objectid > logical ||
1500             found_key->objectid + size <= logical) {
1501                 pr_debug("logical %llu is not within any extent\n", logical);
1502                 return -ENOENT;
1503         }
1504
1505         eb = path->nodes[0];
1506         item_size = btrfs_item_size_nr(eb, path->slots[0]);
1507         BUG_ON(item_size < sizeof(*ei));
1508
1509         ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
1510         flags = btrfs_extent_flags(eb, ei);
1511
1512         pr_debug("logical %llu is at position %llu within the extent (%llu "
1513                  "EXTENT_ITEM %llu) flags %#llx size %u\n",
1514                  logical, logical - found_key->objectid, found_key->objectid,
1515                  found_key->offset, flags, item_size);
1516
1517         WARN_ON(!flags_ret);
1518         if (flags_ret) {
1519                 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
1520                         *flags_ret = BTRFS_EXTENT_FLAG_TREE_BLOCK;
1521                 else if (flags & BTRFS_EXTENT_FLAG_DATA)
1522                         *flags_ret = BTRFS_EXTENT_FLAG_DATA;
1523                 else
1524                         BUG_ON(1);
1525                 return 0;
1526         }
1527
1528         return -EIO;
1529 }
1530
1531 /*
1532  * helper function to iterate extent inline refs. ptr must point to a 0 value
1533  * for the first call and may be modified. it is used to track state.
1534  * if more refs exist, 0 is returned and the next call to
1535  * __get_extent_inline_ref must pass the modified ptr parameter to get the
1536  * next ref. after the last ref was processed, 1 is returned.
1537  * returns <0 on error
1538  */
1539 static int __get_extent_inline_ref(unsigned long *ptr, struct extent_buffer *eb,
1540                                    struct btrfs_key *key,
1541                                    struct btrfs_extent_item *ei, u32 item_size,
1542                                    struct btrfs_extent_inline_ref **out_eiref,
1543                                    int *out_type)
1544 {
1545         unsigned long end;
1546         u64 flags;
1547         struct btrfs_tree_block_info *info;
1548
1549         if (!*ptr) {
1550                 /* first call */
1551                 flags = btrfs_extent_flags(eb, ei);
1552                 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
1553                         if (key->type == BTRFS_METADATA_ITEM_KEY) {
1554                                 /* a skinny metadata extent */
1555                                 *out_eiref =
1556                                      (struct btrfs_extent_inline_ref *)(ei + 1);
1557                         } else {
1558                                 WARN_ON(key->type != BTRFS_EXTENT_ITEM_KEY);
1559                                 info = (struct btrfs_tree_block_info *)(ei + 1);
1560                                 *out_eiref =
1561                                    (struct btrfs_extent_inline_ref *)(info + 1);
1562                         }
1563                 } else {
1564                         *out_eiref = (struct btrfs_extent_inline_ref *)(ei + 1);
1565                 }
1566                 *ptr = (unsigned long)*out_eiref;
1567                 if ((unsigned long)(*ptr) >= (unsigned long)ei + item_size)
1568                         return -ENOENT;
1569         }
1570
1571         end = (unsigned long)ei + item_size;
1572         *out_eiref = (struct btrfs_extent_inline_ref *)(*ptr);
1573         *out_type = btrfs_extent_inline_ref_type(eb, *out_eiref);
1574
1575         *ptr += btrfs_extent_inline_ref_size(*out_type);
1576         WARN_ON(*ptr > end);
1577         if (*ptr == end)
1578                 return 1; /* last */
1579
1580         return 0;
1581 }
1582
1583 /*
1584  * reads the tree block backref for an extent. tree level and root are returned
1585  * through out_level and out_root. ptr must point to a 0 value for the first
1586  * call and may be modified (see __get_extent_inline_ref comment).
1587  * returns 0 if data was provided, 1 if there was no more data to provide or
1588  * <0 on error.
1589  */
1590 int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb,
1591                             struct btrfs_key *key, struct btrfs_extent_item *ei,
1592                             u32 item_size, u64 *out_root, u8 *out_level)
1593 {
1594         int ret;
1595         int type;
1596         struct btrfs_extent_inline_ref *eiref;
1597
1598         if (*ptr == (unsigned long)-1)
1599                 return 1;
1600
1601         while (1) {
1602                 ret = __get_extent_inline_ref(ptr, eb, key, ei, item_size,
1603                                               &eiref, &type);
1604                 if (ret < 0)
1605                         return ret;
1606
1607                 if (type == BTRFS_TREE_BLOCK_REF_KEY ||
1608                     type == BTRFS_SHARED_BLOCK_REF_KEY)
1609                         break;
1610
1611                 if (ret == 1)
1612                         return 1;
1613         }
1614
1615         /* we can treat both ref types equally here */
1616         *out_root = btrfs_extent_inline_ref_offset(eb, eiref);
1617
1618         if (key->type == BTRFS_EXTENT_ITEM_KEY) {
1619                 struct btrfs_tree_block_info *info;
1620
1621                 info = (struct btrfs_tree_block_info *)(ei + 1);
1622                 *out_level = btrfs_tree_block_level(eb, info);
1623         } else {
1624                 ASSERT(key->type == BTRFS_METADATA_ITEM_KEY);
1625                 *out_level = (u8)key->offset;
1626         }
1627
1628         if (ret == 1)
1629                 *ptr = (unsigned long)-1;
1630
1631         return 0;
1632 }
1633
1634 static int iterate_leaf_refs(struct extent_inode_elem *inode_list,
1635                                 u64 root, u64 extent_item_objectid,
1636                                 iterate_extent_inodes_t *iterate, void *ctx)
1637 {
1638         struct extent_inode_elem *eie;
1639         int ret = 0;
1640
1641         for (eie = inode_list; eie; eie = eie->next) {
1642                 pr_debug("ref for %llu resolved, key (%llu EXTEND_DATA %llu), "
1643                          "root %llu\n", extent_item_objectid,
1644                          eie->inum, eie->offset, root);
1645                 ret = iterate(eie->inum, eie->offset, root, ctx);
1646                 if (ret) {
1647                         pr_debug("stopping iteration for %llu due to ret=%d\n",
1648                                  extent_item_objectid, ret);
1649                         break;
1650                 }
1651         }
1652
1653         return ret;
1654 }
1655
1656 /*
1657  * calls iterate() for every inode that references the extent identified by
1658  * the given parameters.
1659  * when the iterator function returns a non-zero value, iteration stops.
1660  */
1661 int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
1662                                 u64 extent_item_objectid, u64 extent_item_pos,
1663                                 int search_commit_root,
1664                                 iterate_extent_inodes_t *iterate, void *ctx)
1665 {
1666         int ret;
1667         struct btrfs_trans_handle *trans = NULL;
1668         struct ulist *refs = NULL;
1669         struct ulist *roots = NULL;
1670         struct ulist_node *ref_node = NULL;
1671         struct ulist_node *root_node = NULL;
1672         struct seq_list tree_mod_seq_elem = SEQ_LIST_INIT(tree_mod_seq_elem);
1673         struct ulist_iterator ref_uiter;
1674         struct ulist_iterator root_uiter;
1675
1676         pr_debug("resolving all inodes for extent %llu\n",
1677                         extent_item_objectid);
1678
1679         if (!search_commit_root) {
1680                 trans = btrfs_join_transaction(fs_info->extent_root);
1681                 if (IS_ERR(trans))
1682                         return PTR_ERR(trans);
1683                 btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem);
1684         } else {
1685                 down_read(&fs_info->commit_root_sem);
1686         }
1687
1688         ret = btrfs_find_all_leafs(trans, fs_info, extent_item_objectid,
1689                                    tree_mod_seq_elem.seq, &refs,
1690                                    &extent_item_pos);
1691         if (ret)
1692                 goto out;
1693
1694         ULIST_ITER_INIT(&ref_uiter);
1695         while (!ret && (ref_node = ulist_next(refs, &ref_uiter))) {
1696                 ret = __btrfs_find_all_roots(trans, fs_info, ref_node->val,
1697                                              tree_mod_seq_elem.seq, &roots);
1698                 if (ret)
1699                         break;
1700                 ULIST_ITER_INIT(&root_uiter);
1701                 while (!ret && (root_node = ulist_next(roots, &root_uiter))) {
1702                         pr_debug("root %llu references leaf %llu, data list "
1703                                  "%#llx\n", root_node->val, ref_node->val,
1704                                  ref_node->aux);
1705                         ret = iterate_leaf_refs((struct extent_inode_elem *)
1706                                                 (uintptr_t)ref_node->aux,
1707                                                 root_node->val,
1708                                                 extent_item_objectid,
1709                                                 iterate, ctx);
1710                 }
1711                 ulist_free(roots);
1712         }
1713
1714         free_leaf_list(refs);
1715 out:
1716         if (!search_commit_root) {
1717                 btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem);
1718                 btrfs_end_transaction(trans, fs_info->extent_root);
1719         } else {
1720                 up_read(&fs_info->commit_root_sem);
1721         }
1722
1723         return ret;
1724 }
1725
1726 int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info,
1727                                 struct btrfs_path *path,
1728                                 iterate_extent_inodes_t *iterate, void *ctx)
1729 {
1730         int ret;
1731         u64 extent_item_pos;
1732         u64 flags = 0;
1733         struct btrfs_key found_key;
1734         int search_commit_root = path->search_commit_root;
1735
1736         ret = extent_from_logical(fs_info, logical, path, &found_key, &flags);
1737         btrfs_release_path(path);
1738         if (ret < 0)
1739                 return ret;
1740         if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
1741                 return -EINVAL;
1742
1743         extent_item_pos = logical - found_key.objectid;
1744         ret = iterate_extent_inodes(fs_info, found_key.objectid,
1745                                         extent_item_pos, search_commit_root,
1746                                         iterate, ctx);
1747
1748         return ret;
1749 }
1750
1751 typedef int (iterate_irefs_t)(u64 parent, u32 name_len, unsigned long name_off,
1752                               struct extent_buffer *eb, void *ctx);
1753
1754 static int iterate_inode_refs(u64 inum, struct btrfs_root *fs_root,
1755                               struct btrfs_path *path,
1756                               iterate_irefs_t *iterate, void *ctx)
1757 {
1758         int ret = 0;
1759         int slot;
1760         u32 cur;
1761         u32 len;
1762         u32 name_len;
1763         u64 parent = 0;
1764         int found = 0;
1765         struct extent_buffer *eb;
1766         struct btrfs_item *item;
1767         struct btrfs_inode_ref *iref;
1768         struct btrfs_key found_key;
1769
1770         while (!ret) {
1771                 ret = btrfs_find_item(fs_root, path, inum,
1772                                 parent ? parent + 1 : 0, BTRFS_INODE_REF_KEY,
1773                                 &found_key);
1774
1775                 if (ret < 0)
1776                         break;
1777                 if (ret) {
1778                         ret = found ? 0 : -ENOENT;
1779                         break;
1780                 }
1781                 ++found;
1782
1783                 parent = found_key.offset;
1784                 slot = path->slots[0];
1785                 eb = btrfs_clone_extent_buffer(path->nodes[0]);
1786                 if (!eb) {
1787                         ret = -ENOMEM;
1788                         break;
1789                 }
1790                 extent_buffer_get(eb);
1791                 btrfs_tree_read_lock(eb);
1792                 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1793                 btrfs_release_path(path);
1794
1795                 item = btrfs_item_nr(slot);
1796                 iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
1797
1798                 for (cur = 0; cur < btrfs_item_size(eb, item); cur += len) {
1799                         name_len = btrfs_inode_ref_name_len(eb, iref);
1800                         /* path must be released before calling iterate()! */
1801                         pr_debug("following ref at offset %u for inode %llu in "
1802                                  "tree %llu\n", cur, found_key.objectid,
1803                                  fs_root->objectid);
1804                         ret = iterate(parent, name_len,
1805                                       (unsigned long)(iref + 1), eb, ctx);
1806                         if (ret)
1807                                 break;
1808                         len = sizeof(*iref) + name_len;
1809                         iref = (struct btrfs_inode_ref *)((char *)iref + len);
1810                 }
1811                 btrfs_tree_read_unlock_blocking(eb);
1812                 free_extent_buffer(eb);
1813         }
1814
1815         btrfs_release_path(path);
1816
1817         return ret;
1818 }
1819
1820 static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root,
1821                                  struct btrfs_path *path,
1822                                  iterate_irefs_t *iterate, void *ctx)
1823 {
1824         int ret;
1825         int slot;
1826         u64 offset = 0;
1827         u64 parent;
1828         int found = 0;
1829         struct extent_buffer *eb;
1830         struct btrfs_inode_extref *extref;
1831         struct extent_buffer *leaf;
1832         u32 item_size;
1833         u32 cur_offset;
1834         unsigned long ptr;
1835
1836         while (1) {
1837                 ret = btrfs_find_one_extref(fs_root, inum, offset, path, &extref,
1838                                             &offset);
1839                 if (ret < 0)
1840                         break;
1841                 if (ret) {
1842                         ret = found ? 0 : -ENOENT;
1843                         break;
1844                 }
1845                 ++found;
1846
1847                 slot = path->slots[0];
1848                 eb = btrfs_clone_extent_buffer(path->nodes[0]);
1849                 if (!eb) {
1850                         ret = -ENOMEM;
1851                         break;
1852                 }
1853                 extent_buffer_get(eb);
1854
1855                 btrfs_tree_read_lock(eb);
1856                 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1857                 btrfs_release_path(path);
1858
1859                 leaf = path->nodes[0];
1860                 item_size = btrfs_item_size_nr(leaf, slot);
1861                 ptr = btrfs_item_ptr_offset(leaf, slot);
1862                 cur_offset = 0;
1863
1864                 while (cur_offset < item_size) {
1865                         u32 name_len;
1866
1867                         extref = (struct btrfs_inode_extref *)(ptr + cur_offset);
1868                         parent = btrfs_inode_extref_parent(eb, extref);
1869                         name_len = btrfs_inode_extref_name_len(eb, extref);
1870                         ret = iterate(parent, name_len,
1871                                       (unsigned long)&extref->name, eb, ctx);
1872                         if (ret)
1873                                 break;
1874
1875                         cur_offset += btrfs_inode_extref_name_len(leaf, extref);
1876                         cur_offset += sizeof(*extref);
1877                 }
1878                 btrfs_tree_read_unlock_blocking(eb);
1879                 free_extent_buffer(eb);
1880
1881                 offset++;
1882         }
1883
1884         btrfs_release_path(path);
1885
1886         return ret;
1887 }
1888
1889 static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
1890                          struct btrfs_path *path, iterate_irefs_t *iterate,
1891                          void *ctx)
1892 {
1893         int ret;
1894         int found_refs = 0;
1895
1896         ret = iterate_inode_refs(inum, fs_root, path, iterate, ctx);
1897         if (!ret)
1898                 ++found_refs;
1899         else if (ret != -ENOENT)
1900                 return ret;
1901
1902         ret = iterate_inode_extrefs(inum, fs_root, path, iterate, ctx);
1903         if (ret == -ENOENT && found_refs)
1904                 return 0;
1905
1906         return ret;
1907 }
1908
1909 /*
1910  * returns 0 if the path could be dumped (probably truncated)
1911  * returns <0 in case of an error
1912  */
1913 static int inode_to_path(u64 inum, u32 name_len, unsigned long name_off,
1914                          struct extent_buffer *eb, void *ctx)
1915 {
1916         struct inode_fs_paths *ipath = ctx;
1917         char *fspath;
1918         char *fspath_min;
1919         int i = ipath->fspath->elem_cnt;
1920         const int s_ptr = sizeof(char *);
1921         u32 bytes_left;
1922
1923         bytes_left = ipath->fspath->bytes_left > s_ptr ?
1924                                         ipath->fspath->bytes_left - s_ptr : 0;
1925
1926         fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr;
1927         fspath = btrfs_ref_to_path(ipath->fs_root, ipath->btrfs_path, name_len,
1928                                    name_off, eb, inum, fspath_min, bytes_left);
1929         if (IS_ERR(fspath))
1930                 return PTR_ERR(fspath);
1931
1932         if (fspath > fspath_min) {
1933                 ipath->fspath->val[i] = (u64)(unsigned long)fspath;
1934                 ++ipath->fspath->elem_cnt;
1935                 ipath->fspath->bytes_left = fspath - fspath_min;
1936         } else {
1937                 ++ipath->fspath->elem_missed;
1938                 ipath->fspath->bytes_missing += fspath_min - fspath;
1939                 ipath->fspath->bytes_left = 0;
1940         }
1941
1942         return 0;
1943 }
1944
1945 /*
1946  * this dumps all file system paths to the inode into the ipath struct, provided
1947  * is has been created large enough. each path is zero-terminated and accessed
1948  * from ipath->fspath->val[i].
1949  * when it returns, there are ipath->fspath->elem_cnt number of paths available
1950  * in ipath->fspath->val[]. when the allocated space wasn't sufficient, the
1951  * number of missed paths in recored in ipath->fspath->elem_missed, otherwise,
1952  * it's zero. ipath->fspath->bytes_missing holds the number of bytes that would
1953  * have been needed to return all paths.
1954  */
1955 int paths_from_inode(u64 inum, struct inode_fs_paths *ipath)
1956 {
1957         return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path,
1958                              inode_to_path, ipath);
1959 }
1960
1961 struct btrfs_data_container *init_data_container(u32 total_bytes)
1962 {
1963         struct btrfs_data_container *data;
1964         size_t alloc_bytes;
1965
1966         alloc_bytes = max_t(size_t, total_bytes, sizeof(*data));
1967         data = vmalloc(alloc_bytes);
1968         if (!data)
1969                 return ERR_PTR(-ENOMEM);
1970
1971         if (total_bytes >= sizeof(*data)) {
1972                 data->bytes_left = total_bytes - sizeof(*data);
1973                 data->bytes_missing = 0;
1974         } else {
1975                 data->bytes_missing = sizeof(*data) - total_bytes;
1976                 data->bytes_left = 0;
1977         }
1978
1979         data->elem_cnt = 0;
1980         data->elem_missed = 0;
1981
1982         return data;
1983 }
1984
1985 /*
1986  * allocates space to return multiple file system paths for an inode.
1987  * total_bytes to allocate are passed, note that space usable for actual path
1988  * information will be total_bytes - sizeof(struct inode_fs_paths).
1989  * the returned pointer must be freed with free_ipath() in the end.
1990  */
1991 struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root,
1992                                         struct btrfs_path *path)
1993 {
1994         struct inode_fs_paths *ifp;
1995         struct btrfs_data_container *fspath;
1996
1997         fspath = init_data_container(total_bytes);
1998         if (IS_ERR(fspath))
1999                 return (void *)fspath;
2000
2001         ifp = kmalloc(sizeof(*ifp), GFP_NOFS);
2002         if (!ifp) {
2003                 kfree(fspath);
2004                 return ERR_PTR(-ENOMEM);
2005         }
2006
2007         ifp->btrfs_path = path;
2008         ifp->fspath = fspath;
2009         ifp->fs_root = fs_root;
2010
2011         return ifp;
2012 }
2013
2014 void free_ipath(struct inode_fs_paths *ipath)
2015 {
2016         if (!ipath)
2017                 return;
2018         vfree(ipath->fspath);
2019         kfree(ipath);
2020 }